| name | fleet-status |
| description | Inspect OpenFleet agent health, tasks, jobs, and live session activity. |
Fleet Status
Use this skill when checking whether the OpenFleet control plane and agents are healthy.
- Start with the top-level dashboard:
node bin/openfleet status.
- Check lifecycle health for session-level problems:
node bin/openfleet lifecycle status.
- Use
node bin/ops summary for a concise fleet snapshot.
- Use
node bin/ops agents to list registered agents and metadata.
- Inspect task backlog with
node bin/ops tasks --json.
- Inspect job state with
node bin/ops jobs --json when work appears stuck.
- Drill into one agent with
node bin/ops inspect <agent>.
- For live message flow, use
node bin/ops tail <agent> or node bin/ops follow <agent>.
- If messaging seems broken, verify sends with
node bin/send <agent> "<message>" and confirm the event lands.
Useful state locations from this repo:
~/.openfleet/agents.json
~/.openfleet/registry.json
~/.openfleet/tasks.json
~/.openfleet/jobs.json
~/.openfleet/events.jsonl
Good diagnosis flow:
- Confirm the fleet is up.
- Identify which agent or task is unhealthy.
- Check recent session output.
- Check queued tasks and jobs.
- Check the event log if state and UI disagree.
Report which agents are healthy, which are stale or broken, and the next operator action needed.